OpenAI will launch its latest AI model, Orion, later this year. This must be up to a hundred times more powerful than GPT-4.

We can expect a little more from OpenAI this year. The Verge has learned from well-informed sources that OpenAI plans to launch its next major LLM in December. Microsoft might even find out about it during the Ignite conference in November. The model has the internal name Orion, but if OpenAI continues to follow its naming scheme it would be the long-awaited GPT-5.

Not much is known about Orion, a name that refers to a galaxy visible in the Northern Hemisphere during winter. According to The Verge, the model is said to be up to a hundred times more powerful than GPT-4. The model’s training was completed in September.

The introduction of the o1 model called Strawberry proved to be the perfect springboard for Orion. OpenAI o1 has a much broader reasoning ability. The model was also used by OpenAI to provide synthetic data for training Orion.

Not for the general public

Unlike previous GPT model launches, Orion would not immediately be made available to the general public via ChatGPT, according to The Verge. Instead, OpenAI plans to initially provide access to companies it works closely with so they can develop their own products and features.

This corresponds to the new course that the OpenAI company is taking. The company is increasingly transforming from a non-profit research organization to a for-profit business enterprise. OpenAI must therefore think about its most lucrative customers and partners first. A $6.6 billion capital injection will reinforce this transformation.

General intelligence

With Orion, OpenAI will take another major step toward the company’s ultimate dream: developing general artificial intelligence on par with the human brain. Not everyone within the organization seems to be equally supportive of these plans.

Miles Brundage, a senior scientist at the company, resigned this week because he said, “The world is far from ready for general artificial intelligence, including OpenAI itself.” He is one of many OpenAI executives to recently leave the company because they are dissatisfied with the company’s current direction, which prioritizes commercial interests over safe and responsible research.